A Novel Contents Delivery Network System for Efficiently Resolving Long-tailed User Requests

摘要

Recently, the delivery frequency of the enormous unpopular High Quality Video contents has been increasing due to the emergence of long-tail phenomena. We propose a new CDN system that uses Surrogate Server (SS) clustering and 2 steps download mechanism. In the proposed system, all contents are classified into three classes based on the number of demands, and different download mechanisms are applied according to their class. The user access time and Original Server processing rate are evaluated by the simulation.